Author

He Zehao, Doctoral Candidate, Renmin University of China Law School.

Abstract

The black–box nature of Artificial Intelligence systems triggers an information cost crisis in torts. In response, the strict liability proposal advocates replacing fault-based liability and strict products liability—both of which involve high information costs—with a pure strict liability regime to improve the deterrent effect of tort law. However, this proposal exaggerates the information costs of fault-based liability, overlooks the necessity of bilateral prevention, fails to account for the uncertainty of defendants' behavior, and may harm innovation due to its anticompetitive effects. In contrast, fault-based liability possesses an underappreciated advantage in information production, as it can improve the functioning of market mechanisms and government regulation through various types of granular information; these tools are often more suitable than tort law itself for incentivizing safety investments by market actors. Even when considering the dual goals of deterrence and compensation, maintaining tort law centered on fault remains a viable transitional solution.

Keywords:information costs;information production;artificial intelligence;liability rules;strict liability;fault-based liability